Machine vision computer optical recognition systems have evolved to become a mainstream automation tool. Accurate placements of components are essential in the manufacture of many electronic systems. The significant challenge for machine vision in automated applications is maintaining the ability to locate reference patterns integrated into the hardware despite changes in material appearance. Normal process variations result in a number of unpredictable and undesirable conditions for optical recognition. Among these are: contrast reversal, intensity gradients, low contrast gradients, angular uncertainties, blur caused by changes in depth of field, physical scale changes, partial obliteration of the reference patterns, and patterns of similar ambiguity within the vision processing field of the view window.
Advanced electronic systems utilize Low Temperature Co-Fired Ceramic (LTCC) substrate technologies in their construction. These technologies make use of an active semiconductor chip mounted to a thermal and electrical conductive material, as well as a heat spreader directly attached to the chassis giving the entire assembly superior thermal and electrical characteristics.
The primary function of machine visioning systems is to distinguish features from the given generated image. The reflective properties of LTCC substrates make proper feature recognition a challenging obstacle. LTCC substrates are manufactured with variations in the shades and colors of the ceramics as well as pad metallurgy appearance. The comparative reflective properties of LTCC substrates and metal pads show many similarities in the degree of light reflection of the two, thus not offering much contrast for feature recognition.
Due to the physical characteristics of LTCC substrates and the manufacturing techniques used in making LTCC substrates, it becomes disadvantageous to utilize printed features on the substrate for machine vision patterns. Low cost gradients, registration challenges, and random shrinkage of the LTCC substrate all contribute to process variations. Also, if the corners of the rectangular through hole served as the vision system pattern reference, that image would change after installation of the heat spreader, thereby further confusing the vision system during the semiconductor place.